012 – Application of Artificial Neural Networks for Classification of Drilling Operations: The deepwater wells case of exploration and production

Valter Chaile, Sergio Moro, Aristides Carneiro and Ricardo F. Ramos
Revista Ibérica de Sistemas e Tecnologias de Informação
Abstract

The application of automatic methods for the classification of unstructured text is precious for the Oil&Gas industry. Drilling is an operation that entails high costs that demands efficiency. A classification of the various operations during drilling is vital to generate assumptions of duration for the design of new
wells. For this paper, two classification analyses for operation classification were conducted to identify the Non-Productive Time (NPT) and Productive Time (PT) best model. Conclusions led to Multi-layer Perceptron (MLP) as the best model. The classification system can produce an accurate and detailed report on the activities performed during the drilling of a well. Through this work, it is possible to
conclude that the currently available daily drilling report represents a rich source of
information and can optimize the oil well construction process.

© 2022 CICEE. Todos os direitos reservados